Applied Education Course Matching Method Based on Behavior Analysis

Abstract

The present invention relates to the technical field of curriculum matching and recommendation, and specifically relates to an application-oriented education curriculum matching method based on behavior analysis. The method includes: obtaining all dimension indicators of a student to be curriculum-matched under each preset dimension in the current time period, and obtaining a set of historical similar students corresponding to the student to be curriculum-matched; determining the current stable relevance of the student to be curriculum-matched under each preset dimension, and obtaining the historical stable relevance of each historical student under each preset dimension; determining the learning stability credibility and the interest learning stability degree under each preset dimension; judging whether the student to be curriculum-matched is a learning-stable student; if the student to be curriculum-matched is not a learning-stable student, then correcting the set of historical similar students, and performing application-oriented education curriculum matching based on the corrected set of historical similar students. The present invention realizes application-oriented education curriculum matching and improves the rationality of recommending application-oriented education curriculum matching for students.

Description

Technical Field

[0001] The present invention relates to the technical field of course matching recommendation, and particularly to an application-oriented education course matching method based on behavior analysis. Background Art

[0002] An application-oriented course refers to an application-oriented course that takes the integration of production and education and school-enterprise cooperation as the teaching mode and emphasizes and highlights the application and practicality. In order to facilitate students' learning of application-oriented education courses, it is often necessary to match and recommend suitable application-oriented education courses for students. Currently, when recommending objects to users, the commonly used method is to recommend objects to users according to the set of similar users stored historically.

[0003] However, when matching and recommending application-oriented education courses for students according to the set of similar students stored historically, the following technical problems often exist:

[0004] Since the learning situation of students often changes dynamically, that is, the learning situation of some students is often unstable. Therefore, the students in the set of similar students stored historically may no longer be the students with similar learning situations to the students waiting for course recommendation at the current stage. At this time, if application-oriented education courses are matched and recommended for students based on the set of similar students stored historically, the rationality of matching and recommending application-oriented education courses for students may be poor. Summary of the Invention

[0005] In order to solve the technical problem of poor rationality in matching and recommending application-oriented education courses for students, the present invention proposes an application-oriented education course matching method based on behavior analysis.

[0006] In a first aspect, the present invention provides an application-oriented education course matching method based on behavior analysis, and the method includes:

[0007] Obtain all dimension indicators of the student waiting for course matching in each preset dimension during the current time period as the current dimension indicators, and obtain the set of historical similar students corresponding to the student waiting for course matching;

[0008] According to the distribution change of all current dimension indicators in each preset dimension, determine the current stable correlation of the student waiting for course matching in each preset dimension. Similarly, obtain the historical stable correlation of each historical student in each preset dimension;

[0009] According to the historical stable correlation of all historical students in each preset dimension, determine the learning stability credibility in each preset dimension;

[0010] According to the current stable correlation and learning stability credibility in each preset dimension, determine the interest learning stability degree of the student waiting for course matching in each preset dimension;

[0011] According to the stability of interest learning of the student to be course-matched under all preset dimensions, determine whether the student to be course-matched is a learning-stable student;

[0012] If the student to be course-matched is a learning-stable student, then perform applied education course matching based on the set of historically similar students;

[0013] If the student to be course-matched is not a learning-stable student, then correct the set of historically similar students, and perform applied education course matching based on the corrected set of historically similar students.

[0014] Combined with the above first aspect, in a possible implementation manner, the determining the current stability correlation of the student to be course-matched under each preset dimension according to the distribution change of all current dimension indicators under each preset dimension includes:

[0015] Construct a current dimension indicator sequence under each preset dimension from all current dimension indicators under each preset dimension;

[0016] Determine the current autocorrelation coefficient of the current dimension indicator sequence under each preset dimension at a preset time interval as the current autocorrelation coefficient of the student to be course-matched under each preset dimension;

[0017] Determine the current stability correlation of the student to be course-matched under each preset dimension according to the current autocorrelation coefficient under each preset dimension and the range of the current dimension indicator sequence thereunder.

[0018] Combined with the above first aspect, in a possible implementation manner, the determining the current stability correlation of the student to be course-matched under each preset dimension according to the current autocorrelation coefficient under each preset dimension and the range of the current dimension indicator sequence thereunder includes:

[0019] Determine the range of the current dimension indicator sequence under each preset dimension as the current range under each preset dimension;

[0020] Determine the current local stability factor under each preset dimension according to the current range under each preset dimension, where the current range and the current local stability factor are negatively correlated;

[0021] Determine the product of the current autocorrelation coefficient and the current local stability factor under each preset dimension as the current stability correlation of the student to be course-matched under each preset dimension.

[0022] Combined with the above first aspect, in a possible implementation manner, the determining the learning stability credibility under each preset dimension according to the historical stability correlation of all historical students under each preset dimension includes:

[0023] Determine the information entropy of the historical stability correlation of all historical students under each preset dimension as the chaos change factor under each preset dimension;

[0024] Determine the learning stability credibility under each preset dimension according to the chaos change factor under each preset dimension, where the chaos change factor is negatively correlated with the learning stability credibility.

[0025] Combined with the above first aspect, in a possible implementation manner, the determining the interest learning stability degree of the students to be course-matched under each preset dimension according to the current stability correlation and the learning stability credibility under each preset dimension includes:

[0026] Normalize the product of the current stability correlation and the learning stability credibility under each preset dimension to obtain the interest learning stability degree of the students to be course-matched under each preset dimension.

[0027] Combined with the above first aspect, in a possible implementation manner, the determining whether the students to be course-matched are learning-stable students according to the interest learning stability degree of the students to be course-matched under all preset dimensions includes:

[0028] Judge whether each preset dimension is the current changing dimension according to the interest learning stability degree of the students to be course-matched under each preset dimension;

[0029] If the proportion of the current changing dimensions in all preset dimensions is greater than the preset changing proportion, it is determined that the students to be course-matched are not learning-stable students;

[0030] If the proportion of the current changing dimensions in all preset dimensions is less than or equal to the preset changing proportion, it is determined that the students to be course-matched are learning-stable students.

[0031] Combined with the above first aspect, in a possible implementation manner, the judging whether each preset dimension is the current changing dimension according to the interest learning stability degree of the students to be course-matched under each preset dimension includes:

[0032] Determine any one of the preset dimensions as the marked dimension. If the interest learning stability degree of the students to be course-matched under the marked dimension is less than or equal to the preset stability threshold, then determine the marked dimension as the current changing dimension.

[0033] Combined with the above first aspect, in a possible implementation manner, the correcting the set of historical similar students includes:

[0034] Obtain all dimension indicators of each historical student under each preset dimension during the historical cycle time period;

[0035] Determine the mean value of all dimension indicators of each historical student under each preset dimension within the historical cycle time period as the dimension representative indicator of each historical student under each preset dimension;

[0036] Determine the mean value of all current dimension indicators of the student to be course-matched under each preset dimension as the dimension representative indicator of the student to be course-matched under each preset dimension;

[0037] Obtain the corrected set of historically similar students based on the dimension representative indicators of all historical students under all preset dimensions and the dimension representative indicators of the student to be course-matched under all preset dimensions.

[0038] Combined with the first aspect above, in a possible implementation manner, the obtaining of the corrected set of historically similar students based on the dimension representative indicators of all historical students under all preset dimensions and the dimension representative indicators of the student to be course-matched under all preset dimensions includes:

[0039] Construct a collaborative filtering matrix based on the dimension representative indicators of all historical students under all preset dimensions and the dimension representative indicators of the student to be course-matched under all preset dimensions as the target collaborative filtering matrix;

[0040] According to the target collaborative filtering matrix, use the collaborative filtering algorithm to screen out the historical students who match the student to be course-matched from all historical students to form the corrected set of historically similar students.

[0041] Combined with the first aspect above, in a possible implementation manner, the preset dimension is a single learning dimension or a fused learning dimension, and the fused learning dimension is the fusion of at least two single learning dimensions.

[0042] In the second aspect, the present invention provides an applied education course matching system based on behavior analysis, and the system includes:

[0043] A data acquisition module, configured to acquire all dimension indicators of the student to be course-matched under each preset dimension within the current time period as the current dimension indicators, and acquire the set of historically similar students corresponding to the student to be course-matched;

[0044] A stable correlation determination module, configured to determine the current stable correlation of the student to be course-matched under each preset dimension according to the distribution change of all current dimension indicators under each preset dimension. Similarly, obtain the historical stable correlation of each historical student under each preset dimension;

[0045] A learning stability credibility determination module, configured to determine the learning stability credibility under each preset dimension according to the historical stable correlation of all historical students under each preset dimension;

[0046] An interest learning stability determination module, configured to determine the interest learning stability degree of a student to be course-matched in each preset dimension according to the current stability correlation and learning stability credibility under each preset dimension;

[0047] A learning stability student judgment module, configured to judge whether a student to be course-matched is a learning stability student according to the interest learning stability degree of the student to be course-matched in all preset dimensions;

[0048] An applied education course matching module, configured to, if the student to be course-matched is a learning stability student, perform applied education course matching based on a set of historical similar students;

[0049] A set correction course matching module, configured to, if the student to be course-matched is not a learning stability student, correct the set of historical similar students and perform applied education course matching based on the corrected set of historical similar students.

[0050] In a third aspect, a server is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method in the first aspect or any possible implementation manner of the first aspect.

[0051] In a fourth aspect, a computer program product is provided, including: computer program code, when the computer program code runs on a computer, enabling the computer to execute the method in the first aspect or any possible implementation manner of the first aspect.

[0052] In a fifth aspect, a computer-readable storage medium is provided, storing computer program code, when the computer program code runs on a computer, enabling the computer to execute the method in the first aspect or any possible implementation manner of the first aspect.

[0053] The present invention has the following beneficial effects:

[0054] The application-oriented education curriculum matching method based on behavior analysis of the present invention realizes application-oriented education curriculum matching, solves the technical problem of poor rationality in recommending application-oriented education curriculum matching for students, and improves the rationality of recommending application-oriented education curriculum matching for students. Compared with recommending application-oriented education curriculum matching for students based on the historically stored set of similar students, the present invention analyzes the dimensional indicators representing different learning behaviors of students, quantifies multiple characteristics related to students' learning situations, such as current stable correlation, historical stable correlation, learning stability credibility, and interest learning stability degree, thereby determining whether the student to be curriculum-matched is a learning-stable student. If the student to be curriculum-matched is not a learning-stable student, the historically similar student set is corrected, which improves the rationality of the subsequent similar students participating in the application-oriented education curriculum matching recommendation to a certain extent, and thus improves the rationality of recommending application-oriented education curriculum matching for students. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0056] Figure 1 It is a flowchart of the application-oriented education curriculum matching method based on behavior analysis of the present invention;

[0057] Figure 2 It is a schematic diagram of the composition structure of the application-oriented education curriculum matching system based on behavior analysis of the present invention;

[0058] Figure 3 It is a schematic diagram of the structure of a computer device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of the technical solutions proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0061] ReferenceFigure 1 , which shows the process of some embodiments of the application-oriented education curriculum matching method based on behavior analysis according to the present invention. The application-oriented education curriculum matching method based on behavior analysis includes the following steps:

[0062] Step S1, obtain all dimension indicators of the student to be curriculum-matched in each preset dimension during the current time period as the current dimension indicators, and obtain the set of historical similar students corresponding to the student to be curriculum-matched.

[0063] Among them, the student to be curriculum-matched is the student for whom application-oriented education curriculum matching recommendations are to be made. The current time period can be a period of time with the current moment as the end moment. The students in the set of historical similar students corresponding to the student to be curriculum-matched can be students who are relatively similar to the student to be curriculum-matched in terms of past learning situations. The preset dimension can be a single pre-set learning dimension or a fused learning dimension. The fused learning dimension can be the fusion of at least two single learning dimensions, that is, it can be the fusion of multiple learning-related dimensions. The single learning dimension can be a single learning-related dimension. The dimension indicator under the preset dimension can be the value corresponding to the dimension data under the preset dimension. That is, if the dimension data is a numerical value, the dimension data can be used as the dimension indicator. If the dimension data is not a numerical value, the dimension data can be converted into a numerical value, and the converted numerical value can be used as the dimension indicator. To a certain extent, the dimension indicator can characterize the learning behavior of the student.

[0064] A single learning dimension can be, but is not limited to: the test score dimension of a single subject, the assignment completion rate dimension of a single subject, the assignment correct rate dimension of a single subject, the learning duration dimension of a single subject, the learning time period dimension of a single subject, the knowledge point mastery dimension of a single subject, the learning resource usage dimension of a single subject, and the course absenteeism rate dimension of a single subject. For example, the dimension index under the test score dimension of a single subject can be the average of all test scores of a single subject within a preset unit time. The preset unit time can be a unit time with a preset corresponding duration less than the duration corresponding to the current time period, and it can be 1 day. The dimension index under the assignment completion rate dimension of a single subject can be the assignment completion rate of a single subject within a preset unit time. The dimension index under the assignment correct rate dimension of a single subject can be the assignment correct rate of a single subject within a preset unit time. The dimension index under the learning duration dimension of a single subject can be the learning duration of a single subject within a preset unit time. The dimension data under the learning time period dimension of a single subject can be morning, afternoon, or evening, that is, it represents that the learning of a single subject by a student within a preset unit time can be in the morning, afternoon, or evening. Since it is not a numerical value, it can be converted into a numerical value. For example, 0000 can be used to represent morning, 0001 can be used to represent afternoon, and 0002 can be used to represent evening. At this time, the dimension index under the learning time period dimension of a single subject can be 0000, 0001, or 0002. The dimension index under the knowledge point mastery dimension of a single subject can be the score given by the teacher for the student's mastery of the subject. The larger the value, the higher the student's mastery of the subject, and its value range can be [0, 10]. The dimension index under the learning resource usage dimension of a single subject can be different numerical values used to represent different learning resource types. Among them, different learning resource types can include: videos, texts, and interactive exercises. The dimension index under the course absenteeism rate dimension of a single subject can be the course absenteeism rate.

[0065] The fused learning dimensions can be, but are not limited to: the learning input dimension of a single subject, the learning outcome dimension of a single subject, and the learning habit dimension of a single subject. The learning input dimension of a single subject can be the fusion of the learning duration dimension and the homework completion rate dimension of a single subject. The learning outcome dimension of a single subject can be the fusion of the test score dimension and the homework accuracy rate dimension of a single subject. The learning habit dimension of a single subject can be the fusion of the learning time period dimension and the learning resource usage dimension of a single subject. The method for obtaining the dimension indicators under the fused learning dimension can be: taking the dimension indicators of all single learning dimensions included in this fused learning dimension for the same student within a preset unit time to form a target vector under this fused learning dimension, classifying all the target vectors under this fused learning dimension, and representing the different categories obtained at this time with different numerical values, and recording the numerical value corresponding to the category to which the target vector belongs as the dimension indicator under the fused learning dimension. Among them, the algorithms for realizing classification can be, but are not limited to: decision tree algorithm, logistic regression algorithm, K-Means clustering algorithm, and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. For example, the method for obtaining the dimension indicators under the learning input dimension of a single subject can include: taking the dimension indicators of all single learning dimensions included in the learning input dimension for the same student within a preset unit time to form a target vector under the learning input dimension, classifying all the target vectors under the learning input dimension through the DBSCAN algorithm, and representing the different categories obtained at this time with different numerical values, and recording the numerical value corresponding to the category to which the target vector under the learning input dimension belongs as the dimension indicator under the learning input dimension.

[0066] It should be noted that the number of preset dimensions can be set in advance. Generally speaking, the more the number of preset dimensions, the relatively more reasonable the students finally screened by the embodiments of the present invention who are similar to the learning situation of the students to be matched with the course, and the relatively more reasonable the applied education courses currently matched for the students to be matched with the course. Secondly, the missing data can be filled with the mean, median or mode. For example, if a student does not take a test for a certain subject on a certain day, the mean of all test scores of this subject for this student within the week when this day is located can be used to represent the average test score of this student on this day, and it can be used to represent the dimension indicator of this student on this day under the test score dimension of this subject. If a student does not use any learning resources to study a certain subject on a certain day, the numerical value corresponding to the type of learning resource most used by this student for this subject within the week when this day is located can be used to represent the dimension indicator of this student on this day under the learning resource usage dimension of this subject.

[0067] As an example, this step can include the following steps:

[0068] First, obtain the dimensional indicators of the students to be course-matched in each preset dimension within each preset unit time; form all the dimensional indicators of the students to be course-matched in the same preset dimension within all preset unit times in the current time period, which are all the dimensional indicators of the students to be course-matched in the same preset dimension in the current time period; and use the dimensional indicators in the current time period as the current dimensional indicators.

[0069] Second, similarly, obtain all the dimensional indicators of the students to be course-matched in each preset dimension within the historical time period, and obtain all the dimensional indicators of each historical student in each preset dimension within the historical cycle time period.

[0070] Among them, the historical time period can be the time period before the current time period, and the end moment of the historical time period can be the start moment of the current time period. The historical cycle time period can include the time stage corresponding to the historical time period and the time stage corresponding to the current time period. For example, if the student to be course-matched is a student in the third grade of primary school and has studied for 6 weeks, the historical time period can be the first 2 weeks when the student to be course-matched studied in the third grade of primary school, the current time period can be the last 4 weeks when the student to be course-matched studied in the third grade of primary school, and the historical cycle time period can be the entire period when the historical student studied in the third grade of primary school.

[0071] Third, determine the overall historical dimensional indicators of the students to be course-matched in each preset dimension as the average value of all the dimensional indicators of the students to be course-matched in each preset dimension within the historical time period.

[0072] Fourth, determine the overall historical dimensional indicators of each historical student in each preset dimension as the average value of all the dimensional indicators of each historical student in each preset dimension within the historical cycle time period.

[0073] Fifth, obtain the set of historical similar students corresponding to the students to be course-matched.

[0074] For example, based on the overall historical dimensional indicators of the students to be course-matched in each preset dimension and the overall historical dimensional indicators of each historical student in each preset dimension, a collaborative filtering matrix can be constructed as the historical collaborative filtering matrix, and based on the historical collaborative filtering matrix, through the collaborative filtering algorithm, historical students with learning situations similar to those of the students to be course-matched can be selected from all historical students to form the set of historical similar students.

[0075] Step S2, determine the current stable correlation of the students to be course-matched in each preset dimension according to the distribution change of all the current dimensional indicators in each preset dimension. Similarly, obtain the historical stable correlation of each historical student in each preset dimension.

[0076] As an example, determining the current stable correlation of students to be course-matched in each preset dimension based on the distribution changes of all current dimension indicators under each preset dimension may include the following steps:

[0077] First, all current dimension indicators under each preset dimension are formed into a current dimension indicator sequence under each preset dimension.

[0078] Among them, the current dimension indicator sequence can be a time series, and the earlier the acquisition time corresponding to the current dimension indicator is in the current dimension indicator sequence.

[0079] Second, the autocorrelation coefficient of the current dimension indicator sequence under each preset dimension at a preset time interval is determined as the current autocorrelation coefficient of the students to be course-matched in each preset dimension.

[0080] Among them, the preset time interval can be a preset time interval, which can be 7 days.

[0081] It should be noted that in actual situations, when the autocorrelation coefficient of a sequence is larger, it often means that the current value is more likely to be highly correlated with past values, and it often means that the change trend of the sequence is more likely to be regular or stable. Therefore, when the current autocorrelation coefficient of a student to be course-matched in a certain preset dimension is larger, it often means that the learning behavior of the student to be course-matched related to this preset dimension is more likely to be stable, and it often means that the learning situation of the student to be course-matched in this preset dimension is more likely to remain stable.

[0082] Third, determining the current stable correlation of the students to be course-matched in each preset dimension based on the current autocorrelation coefficient and the range of the current dimension indicator sequence under each preset dimension may include the following sub-steps:

[0083] The first sub-step is to determine the range of the current dimension indicator sequence under each preset dimension as the current range under each preset dimension.

[0084] The second sub-step is to determine the current local stability factor under each preset dimension according to the current range under each preset dimension.

[0085] Among them, the current range may have a negative correlation with the current local stability factor.

[0086] The third sub-step is to determine the product of the current autocorrelation coefficient and the current local stability factor under each preset dimension as the current stable correlation of the students to be course-matched in each preset dimension.

[0087] For example, the formula corresponding to determining the current stable correlation of a student to be course-matched in a preset dimension may be:

[0088] ; where is the current stable correlation of the student to be course-matched under the i-th preset dimension. i is the serial number of the preset dimension. is the current autocorrelation coefficient of the student to be course-matched under the i-th preset dimension. is the maximum value in the current dimension index sequence under the i-th preset dimension. is the minimum value in the current dimension index sequence under the i-th preset dimension. is the natural exponential function. is the current range under the i-th preset dimension. is the current local stability factor under the i-th preset dimension.

[0089] It should be noted that when is larger, it often indicates that the learning behavior of the student to be course-matched related to the i-th preset dimension is more likely to be stable, and it often indicates that the learning situation of the student to be course-matched under the i-th preset dimension is more likely to remain stable. When is smaller, it often indicates that the fluctuation of the dimension index of the student to be course-matched under the i-th preset dimension is relatively smaller, it often indicates that the learning situation of the student to be course-matched under the i-th preset dimension is more likely to change smoothly, and it often indicates that the learning situation of the student to be course-matched under the i-th preset dimension is more likely to remain stable. Therefore, when is larger, it often indicates that the learning situation of the student to be course-matched under the i-th preset dimension is more likely to remain stable, and it often indicates that the learning behavior under the i-th preset dimension is more likely to have not changed significantly.

[0090] As another example, obtaining the historical stable correlation of each historical student under each preset dimension may include the following steps:

[0091] First step, all the dimension indexes of each historical student under each preset dimension within the historical cycle time period are used to form the historical dimension index sequence of each historical student under each preset dimension.

[0092] Among them, the historical dimension index sequence can be a time series, and the earlier the dimension index is in the historical dimension index sequence, the earlier the collection time corresponding to the dimension index can be.

[0093] Second step, the autocorrelation coefficient of the historical dimension index sequence of each historical student under each preset dimension at a preset time interval is determined as the historical autocorrelation coefficient of each historical student under each preset dimension.

[0094] The third step, determining the historical stability correlation of each historical student in each preset dimension based on the historical autocorrelation coefficient of each historical student in each preset dimension and the range of the historical dimension index sequence thereunder may include the following sub-steps:

[0095] The first sub-step is to determine the historical range of each historical student in each preset dimension as the range of the historical dimension index sequence of each historical student in each preset dimension.

[0096] The second sub-step is to determine the historical local stability factor of each historical student in each preset dimension according to the historical range of each historical student in each preset dimension.

[0097] Among them, the historical range may have a negative correlation with the historical local stability factor.

[0098] The third sub-step is to determine the historical stability correlation of each historical student in each preset dimension as the product of the historical autocorrelation coefficient and the historical local stability factor of each historical student in each preset dimension.

[0099] Step S3, determining the learning stability credibility of each preset dimension according to the historical stability correlation of all historical students in each preset dimension.

[0100] As an example, this step may include the following steps:

[0101] The first step is to determine the chaos change factor of each preset dimension as the information entropy of the historical stability correlation of all historical students in each preset dimension.

[0102] The second step is to determine the learning stability credibility of each preset dimension according to the chaos change factor of each preset dimension.

[0103] Among them, the chaos change factor may have a negative correlation with the learning stability credibility.

[0104] For example, the formula for determining the learning stability credibility corresponding to the preset dimension may be:

[0105] ; where is the learning stability credibility of the i-th preset dimension. i is the serial number of the preset dimension. is the chaos change factor of the i-th preset dimension.

[0106] It should be noted that when is larger, it often indicates that the learning behaviors of different students in the i-th preset dimension are more likely to change. Therefore, when The larger it is, it often indicates that the learning behaviors of different students under the i-th preset dimension are less likely to change, often indicating that the learning behaviors under the i-th preset dimension are more likely to be relatively common learning behaviors, and often indicating that the learning behaviors under the i-th preset dimension are more generally stable.

[0107] Step S4: Determine the degree of interest learning stability of the students to be course-matched under each preset dimension according to the current stability correlation and learning stability credibility under each preset dimension.

[0108] As an example, the product of the current stability correlation and learning stability credibility under each preset dimension can be normalized to obtain the degree of interest learning stability of the students to be course-matched under each preset dimension.

[0109] For example, the formula for determining the degree of interest learning stability corresponding to the students to be course-matched under the preset dimension can be:

[0110] ; where is the degree of interest learning stability of the students to be course-matched under the i-th preset dimension. i is the serial number of the preset dimension. is the normalization function. is the current stability correlation of the students to be course-matched under the i-th preset dimension. is the learning stability credibility under the i-th preset dimension.

[0111] It should be noted that can be used as 's weight. When is larger, it often indicates that the learning situation of the students to be course-matched under the i-th preset dimension is more likely to remain stable, and often indicates that the learning behaviors under the i-th preset dimension are less likely to have changed significantly. When is larger, it often indicates that the learning behaviors of different students under the i-th preset dimension are less likely to change, and often indicates that the learning behaviors under the i-th preset dimension are more generally stable. Therefore, when is larger, it often indicates that the learning situation of the students to be course-matched under the i-th preset dimension is relatively more stable.

[0112] Step S5: Judge whether the students to be course-matched are learning-stable students according to the degree of interest learning stability of the students to be course-matched under all preset dimensions.

[0113] As an example, this step may include the following steps:

[0114] The first step: Judge whether each preset dimension is the current changing dimension according to the degree of interest learning stability of the students to be course-matched under each preset dimension.

[0115] For example, any one of the preset dimensions can be determined as the marked dimension. If the interest learning stability degree of the student to be course-matched under the marked dimension is less than or equal to the preset stability threshold, then the marked dimension is determined as the current variable dimension. Among them, the preset stability threshold can be a threshold set in advance, and it can be 0.5.

[0116] It should be noted that the current variable dimension can represent the preset dimension in which the learning behavior of the student to be course-matched changes greatly.

[0117] In the second step, if the proportion of the current variable dimension among all the preset dimensions is greater than the preset variable proportion, it is determined that the student to be course-matched is not a learning-stable student.

[0118] Among them, the proportion of the current variable dimension among all the preset dimensions can be expressed as: ; where q is the number of current variable dimensions. Q is the number of preset dimensions. The preset variable proportion can be a proportion set in advance, and it can be 0.4.

[0119] It should be noted that when there are more current variable dimensions, it often means that there are more preset dimensions representing large changes in the learning behavior of the student to be course-matched, often indicating that there may be more learning behaviors of the student to be course-matched that have changed, often indicating that the learning situations of the students in the historically stored set of similar students are more likely to be less similar to the current learning situation of the student to be course-matched, and often indicating that when making course recommendations for the above-mentioned student to be course-matched, it is more necessary to correct the set of historically similar students.

[0120] In the third step, if the proportion of the current variable dimension among all the preset dimensions is less than or equal to the preset variable proportion, it is determined that the student to be course-matched is a learning-stable student.

[0121] Step S6, if the student to be course-matched is a learning-stable student, then perform applied education course matching based on the set of historically similar students.

[0122] As an example, if the student to be course-matched is a learning-stable student, then the applied education courses of the historical students in the set of historically similar students can be recommended to the student to be course-matched.

[0123] Step S7, if the student to be course-matched is not a learning-stable student, then correct the set of historically similar students and perform applied education course matching based on the corrected set of historically similar students.

[0124] As an example, this step can include the following steps:

[0125] In the first step, obtain all the dimension indicators of each historical student under each preset dimension during the historical cycle time period.

[0126] In the second step, the mean value of all dimension indicators of each historical student under each preset dimension within the historical cycle time period is determined as the dimension representative indicator of each historical student under each preset dimension.

[0127] In the third step, the mean value of all current dimension indicators of the student to be course-matched under each preset dimension is determined as the dimension representative indicator of the student to be course-matched under each preset dimension.

[0128] In the fourth step, based on the dimension representative indicators of all historical students under all preset dimensions and the dimension representative indicators of the student to be course-matched under all preset dimensions, a corrected set of historically similar students is obtained.

[0129] For example, obtaining the corrected set of historically similar students may include the following sub-steps:

[0130] In the first sub-step, based on the dimension representative indicators of all historical students under all preset dimensions and the dimension representative indicators of the student to be course-matched under all preset dimensions, a collaborative filtering matrix is constructed as the target collaborative filtering matrix.

[0131] In the second sub-step, based on the target collaborative filtering matrix, through the collaborative filtering algorithm, historical students who match the student to be course-matched are selected from all historical students to form a corrected set of historically similar students.

[0132] Among them, the historical students who match the student to be course-matched can be students with a learning situation similar to that of the student to be course-matched currently.

[0133] Optionally, obtaining the corrected set of historically similar students may also include the following sub-steps:

[0134] In the first sub-step, the dimension representative indicators of each historical student under all preset dimensions are used to form a dimension overall vector corresponding to each historical student.

[0135] In the second sub-step, the dimension representative indicators of the student to be course-matched under all preset dimensions are used to form a dimension overall vector corresponding to the student to be course-matched.

[0136] In the third sub-step, clustering is performed on all dimension overall vectors, and the clustering clusters obtained at this time are used as the target clustering clusters.

[0137] In the fourth sub-step, the historical students corresponding to all dimension overall vectors within the target clustering cluster to which the dimension overall vector corresponding to the student to be course-matched belongs are used to form a corrected set of historically similar students.

[0138] In the fifth step, application-oriented education course matching is performed based on the corrected set of historically similar students.

[0139] For example, the applied education courses of the historical students in the corrected historical similar student set can be recommended to the students to be course-matched, so as to realize the matching of applied education courses based on the corrected historical similar student set.

[0140] Reference Figure 2 , based on the same inventive concept as the above method embodiments, the present invention provides an applied education course matching system based on behavior analysis. The system includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of an applied education course matching method based on behavior analysis, which may specifically include:

[0141] A data acquisition module 201, configured to acquire all dimension indicators of the student to be course-matched in each preset dimension within the current time period as the current dimension indicators, and acquire the set of historical similar students corresponding to the student to be course-matched;

[0142] A stable correlation determination module 202, configured to determine the current stable correlation of the student to be course-matched in each preset dimension according to the distribution change of all current dimension indicators in each preset dimension. Similarly, the historical stable correlation of each historical student in each preset dimension is acquired;

[0143] A learning stability credibility determination module 203, configured to determine the learning stability credibility in each preset dimension according to the historical stable correlations of all historical students in each preset dimension;

[0144] An interest learning stability degree determination module 204, configured to determine the interest learning stability degree of the student to be course-matched in each preset dimension according to the current stable correlation and the learning stability credibility in each preset dimension;

[0145] A learning stable student determination module 205, configured to determine whether the student to be course-matched is a learning stable student according to the interest learning stability degrees of the student to be course-matched in all preset dimensions;

[0146] An applied education course matching module 206, configured to, if the student to be course-matched is a learning stable student, perform matching of applied education courses based on the set of historical similar students;

[0147] A set correction course matching module 207, configured to, if the student to be course-matched is not a learning stable student, correct the set of historical similar students, and perform matching of applied education courses based on the corrected set of historical similar students.

[0148] Figure 3 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention. Exemplarily, asFigure 3 As shown in Figure 3 , the computer device 300 includes: a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any one of the above-described application-oriented education curriculum matching methods based on behavior analysis.

[0149] Based on the same inventive concept as the above method embodiments, the present invention provides a server, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes any one of the above application-oriented education curriculum matching methods based on behavior analysis.

[0150] Based on the same inventive concept as the above method embodiments, the present invention provides a computer program product, which includes: computer program code. When the computer program code runs on a computer, the computer is enabled to execute any one of the above application-oriented education curriculum matching methods based on behavior analysis.

[0151] Based on the same inventive concept as the above method embodiments, the present invention provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer is enabled to execute any one of the above application-oriented education curriculum matching methods based on behavior analysis.

[0152] In summary, compared with recommending application-oriented education curriculum matching for students according to the historically stored set of similar students, the present invention analyzes the dimensional indicators representing different learning behaviors of students, quantifies multiple features related to students' learning situations, such as current stability correlation, historical stability correlation, learning stability credibility, and interest learning stability degree, thereby determining whether the student to be curriculum-matched is a learning-stable student. If the student to be curriculum-matched is not a learning-stable student, the historically similar student set is corrected, which improves the rationality of the similar students participating in the subsequent application-oriented education curriculum matching recommendation to a certain extent, and thus improves the rationality of recommending application-oriented education curriculum matching for students.

[0153] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for matching applied education courses based on behavioral analysis, characterized in that: The following steps are involved: Obtain all dimension indicators of the students to be matched with the course under each preset dimension in the current time period as the current dimension indicators, and obtain the set of historically similar students corresponding to the students to be matched with the course; According to the distribution changes of all current dimension indicators under each preset dimension, the current stable correlation of the students to be matched with the course under each preset dimension is determined, including: All current dimension indicators under each preset dimension constitute a current dimension indicator sequence under each preset dimension; the autocorrelation coefficient of the current dimension indicator sequence under each preset dimension at a preset time interval is determined as the current autocorrelation coefficient of the student to be matched with the course under each preset dimension; according to the current autocorrelation coefficient under each preset dimension and the range of the current dimension indicator sequence under it, the current stable correlation of the student to be matched with the course under each preset dimension is determined, including: determining the range of the current dimension indicator sequence under each preset dimension as the current range under each preset dimension; according to the current range under each preset dimension, determining the current local stability factor under each preset dimension, wherein the current range is negatively correlated with the current local stability factor; the product of the current autocorrelation coefficient under each preset dimension and the current local stability factor is determined as the current stable correlation of the student to be matched with the course under each preset dimension; Similarly, obtain the historical stability correlation of each history student in each preset dimension; Based on the historical stability correlation of all history students in each preset dimension, the learning stability credibility under each preset dimension is determined, including: The information entropy of the historical stability correlation of all history students in each preset dimension is determined as the chaotic change factor in each preset dimension; according to the chaotic change factor in each preset dimension, the learning stability credibility in each preset dimension is determined, wherein the chaotic change factor is negatively correlated with the learning stability credibility; According to the current stable relevance and learning stability credibility under each preset dimension, determine the interest learning stability of the students to be matched with the course under each preset dimension; According to the interest and learning stability of the students to be matched with courses in all preset dimensions, it is judged whether the students to be matched with courses are stable students; If the student to be matched with a course is a stable student, then the applied education course matching is performed based on a set of historically similar students; If the student to be matched with a course is not a stable student, the set of historically similar students is revised, and applied education course matching is performed based on the revised set of historically similar students.

2. According to claim 1, a method for matching applied education courses based on behavioral analysis is characterized in that: Determining the interest learning stability of the students to be matched with the course in each preset dimension according to the current stable correlation and learning stability credibility in each preset dimension includes: The product of the current stability correlation and the learning stability credibility under each preset dimension is normalized to obtain the interest learning stability of the students to be matched with the course under each preset dimension.

3. The method for matching applied education courses based on behavioral analysis according to claim 1, characterized in that: The step of judging whether the student to be matched with a course is a stable student in terms of the stability of the student's interest in all preset dimensions includes: According to the stability of the interest learning of the students to be matched with the course in each preset dimension, determine whether each preset dimension is the current variable dimension; If the proportion of the current change dimension in all preset dimensions is greater than the preset change proportion, it is determined that the student to be matched with the course is not a stable student; If the proportion of the current change dimension in all preset dimensions is less than or equal to the preset change proportion, the student to be matched with the course is determined to be a stable learning student.

4. The method for matching applied education courses based on behavioral analysis according to claim 3 is characterized in that: The step of judging whether each preset dimension is a current variable dimension according to the interest learning stability of the students to be matched with the course in each preset dimension includes: Any preset dimension is determined as a marked dimension. If the degree of interest learning stability of the students to be matched with the course under the marked dimension is less than or equal to a preset stability threshold, the marked dimension is determined as the current variable dimension.

5. The method for matching applied education courses based on behavioral analysis according to claim 1, characterized in that: The correction of the historically similar student set includes: Get all dimension indicators of each history student under each preset dimension in the historical period time period; The mean of all dimension indicators of each history student under each preset dimension during the historical period is determined as the dimension representative indicator of each history student under each preset dimension; The mean of all current dimension indicators of the students to be matched with the course in each preset dimension is determined as the dimension representative indicator of the students to be matched with the course in each preset dimension; According to the dimension representative indicators of all historical students in all preset dimensions and the dimension representative indicators of the students to be matched with the courses in all preset dimensions, a revised set of historical similar students is obtained.

6. The method for matching applied education courses based on behavioral analysis according to claim 5, characterized in that: The method of obtaining a corrected set of historical similar students according to the dimension representative indicators of all historical students in all preset dimensions and the dimension representative indicators of the students to be matched with the course in all preset dimensions includes: According to the dimension representative indicators of all historical students in all preset dimensions and the dimension representative indicators of the students to be matched with the courses in all preset dimensions, a collaborative filtering matrix is ​​constructed as the target collaborative filtering matrix; According to the target collaborative filtering matrix, the collaborative filtering algorithm is used to screen out historical students who match the students to be matched with the course from all historical students to form a revised set of similar historical students.

7. The method for matching applied education courses based on behavioral analysis according to claim 1, characterized in that: The preset dimension is a single learning dimension or a fusion learning dimension, and the fusion learning dimension is the fusion of at least two single learning dimensions.

Images